From Notebook to Kubeflow Pipelines with HP Tuning - A Data Science Journey

Explore a comprehensive tutorial on transforming Jupyter Notebooks into scalable Kubeflow Pipelines with hyperparameter tuning. Learn to deploy Kubeflow, convert ML code into composable workflows, ensure reproducibility through immutable snapshots, debug historical versions, and distribute computations. Discover how to leverage Kubeflow components like Pipelines, Kale, Katib, and Snapshot Store to effortlessly scale up machine learning projects. Gain insights into improved development workflows, data management, and the Kubeflow Marketplace. Follow along as the presenters demonstrate cloning repositories, scaling notebooks, and running experiments, all without the need for specific SDKs or CLI commands.

Introduction
What you will learn
Overview
Pipelines
Notebook to Pipeline
Improved Development Workflow
Hyperparameter Tuning
Agenda
Data Management
Kubeflow Marketplace
Cloning the K repository
Scaling
Notebook
Experiments
Summary